Positive Train Control and High Speed Rail

[NellieBly]

The subject of Positive Train Control has been "under the radar" for most railfans, but a noteworthy event happened on July 21. FRA issued its proposed rule requiring railroads to install PTC on all "main lines", all those that carry passenger trains, all those that carry more than five MGT of traffic, and all those that carry any TIH (toxic inhalation hazard) traffic. Implementation plans are due by next April (2010) and systems must be in place by 2015. This is real, folks.

There will be a hearing in Washington on August 13, for any of you who would like to comment (first you have to read the 304-page document describing the proposed rule).

PTC will permit operation of passenger trains at more than 79 MPH, subject to other requirements such as track safety standards. It will positively enforce train separation, and it will permit the removal of obsolete cab signal and intermittent train stop systems. Get ready for the new world!

I just finished wading through the rule in all its detail. The USDOT Office of Policy will be offering appropriate comments in due time, but all told I think FRA did a very good job of laying out the case for PTC, the costs, and the potential benefits for the freight railroads.

Some consider it an "unfunded mandate". The rule, however, makes it pretty clear that the intent is that PTC will replace existing control systems in time. In that case, the net cost to the industry will be quite small, and the potential benefits very large. And we may see lots more higher-speed passenger trains.

[David Benton]

a very positive step forward i think .

[Vincent]

I believe the links to the documents are found on this page.

[orulz]

I don't have the time to wade through a 304 page document. This has the potential to be very helpful to the rail industry in the US in general.

I note the requirement to upgrade many mainline railroads by 2015; but is there anything in the document that specifies a particular technology to use, or at least a stipulation that systems be interoperable?

If all the railroads in the US are implementing this basically from scratch, then they better be planning to implement compatible technology.

[justalurker66]

I've got to read the document ... but 2015 gives the railroads a few more years (I thought the deadline was 2012?) and previous reading makes me thing that "positive train control" may not be as controlling as one may think. What it actually is will vary from railroad to railroad ... although compatibility will need to be maintained for run through power.

Time to go reading ...

[george matthews]

The subject of Positive Train Control has been "under the radar" for most railfans, but a noteworthy event happened on July 21. FRA issued its proposed rule requiring railroads to install PTC on all "main lines", all those that carry passenger trains, all those that carry more than five MGT of traffic, and all those that carry any TIH (toxic inhalation hazard) traffic. Implementation plans are due by next April (2010) and systems must be in place by 2015. This is real, folks.

There will be a hearing in Washington on August 13, for any of you who would like to comment (first you have to read the 304-page document describing the proposed rule).

PTC will permit operation of passenger trains at more than 79 MPH, subject to other requirements such as track safety standards. It will positively enforce train separation, and it will permit the removal of obsolete cab signal and intermittent train stop systems. Get ready for the new world!

I just finished wading through the rule in all its detail. The USDOT Office of Policy will be offering appropriate comments in due time, but all told I think FRA did a very good job of laying out the case for PTC, the costs, and the potential benefits for the freight railroads.

Some consider it an "unfunded mandate". The rule, however, makes it pretty clear that the intent is that PTC will replace existing control systems in time. In that case, the net cost to the industry will be quite small, and the potential benefits very large. And we may see lots more higher-speed passenger trains.

How does it relate to the European ERTMS rules? These are intended to harmonise the various national systems in use and have different layers according to the density of traffic.

[tarheelman]

Thanks for the info, Ms. Bly. This is good news---hopefully, it'll result in faster passenger service outside of the NEC.

[obienick]

Great info!

One question ... who's going to pay?

[Nasadowsk]

Great info!

One question ... who's going to pay?

The railroads.

[Jersey_Mike]

Some consider it an "unfunded mandate". The rule, however, makes it pretty clear that the intent is that PTC will replace existing control systems in time. In that case, the net cost to the industry will be quite small, and the potential benefits very large. And we may see lots more higher-speed passenger trains.

This has all the makings of a disaster in the same vein as the CBTC on SEPTA's trolley tunnel or what happened on the NYC Subway since the Willy B wreck. In both cases the push for "safety" has drastically slowed down both systems. The part of the regulations that really matters is what fulfills the requirements of a PTC system. Until now PTC only meant enforced stops at STOP signals. If a PTC system has to be 100% failsafe we will be looking at an unmitigated disaster as mile long freight trains will be slowed to a crawl to obey some overly conservative speed target or worse get stopped completely because they lose contact with whatever fancy pants GPS based wireless system got installed.

Right now METRA is using a GPS based system to enforce speeds and signals, but it is not 100% accurate on location and trains aren't dumped to 20mph if they lose contact. If this sort of backstop to the traditional signaling system is acceptable PTC then things should be alright, but trying to move trains using wireless technology with 6 8s of service out in some West Virginian gorge is just asking to have our natural rail network brought down to the level of CSX with 22mph intermodals and Amtrak trains running 4 hours late.

PTC will permit operation of passenger trains at more than 79 MPH, subject to other requirements such as track safety standards. It will positively enforce train separation, and it will permit the removal of obsolete cab signal and intermittent train stop systems. Get ready for the new world!

Class 4+ track maintenance requirements is the biggest hurdle to high speed rail. Amtrak could go faster than 80 on any of the hundreds of miles of Pulse Code Cab Signaled lines it runs on, but it doesn't because of the track class and other wear and tear related reasons. Also Amtrak's ACSES uses the existing cab signaling system to control speed down to 20mph where the PTC kicks in at Stop signals.

How does it relate to the European ERTMS rules? These are intended to harmonise the various national systems in use and have different layers according to the density of traffic.

It doesn't. The FRA is not going to regulate any single technology for this. Railroads will be free to choose the system that suits their needs. Don't expect railroads to install anything that requires them to re-signal or re-interlock their existing routes. It costs millions and takes months to install a single interlocking so the PTC systems will most likley be overlays with something like the METRA GPS system being prefered if it fits the definition of PTC. Other than that you'll get things with transponders giving diving distance to speed targets.

[NellieBly]

Let me offer one more clarification. Jersey Mike, there are two very different philosophies here. CBTC, ETMS, and ACSES, although "communications based", are conceptually the same as conventional cab signaling. Virtually all communication is between wayside devices and the train, and is largely limited to transmission of target speeds.

PTC, on the other hand, is vehicle-centric. Each train navigates itself, using GPS and other means (accelerometers, odometer) for location, and relying on movement authorities transmitted from a central office. That's all the central office transmits -- limits to movement authorities. The on-train computer regulates speed based both on the distance to the authority limit and on a database of track geometry, profile, and speeds that is maintained in memory. This allows for "virtual block" operation and, implemented properly, will minimize distance between trains subject to safety constraints.

FRA's rule is not prescriptive. It describes what the system must do, not how it should accomplish it. Each railroad must file a "case" with FRA outlining how the requirements are to be met, and must demonstrate that any proposed system is "at least as safe as" current technology. FRA emphasizes the need for compatibility between railroads, and spends considerable time on how unequipped or non-communicating trains are to be handled. Read the whole thing, and try to put the notion of conventional cab signals out of your mind. That's not how this will work.

[Vincent]

Reading through the pile of ARRA applications it also appears that a large percentage of the cost will be borne by the States that are looking to upgrade existing rail lines to HSR. That could be a big reason why the host railroads suddenly seem to be willing to play ball with Amtrak and HSR. On lines without, the Class Ones will have to pay the entire cost of upgrading the signals. I wonder if UP is secretly talking to Amtrak about a HSR version of the Sunset Limited?

[Jersey_Mike]

PTC, on the other hand, is vehicle-centric. Each train navigates itself, using GPS and other means (accelerometers, odometer) for location, and relying on movement authorities transmitted from a central office. That's all the central office transmits -- limits to movement authorities. The on-train computer regulates speed based both on the distance to the authority limit and on a database of track geometry, profile, and speeds that is maintained in memory. This allows for "virtual block" operation and, implemented properly, will minimize distance between trains subject to safety constraints.

You might see such a system on currently unsignaled lines, but on anything resembling a main line with high traffic densities is going to need something more reliable, ideally with continuous reliable communications of block state. Railroads need to go with something reliable like the French TMV or KVB that either uses coded track circuits or track mounted data beacons (oh that's what ACSES is). At the end of the day the problem will be that a computer model will not run a train as fast as an engineers judgment, either because of conservative margins for error or intermittent updates. Also forget about your Amtrak engineer ever being able to make up time through some aggressive train handling like braking late on an approach.

Read the whole thing, and try to put the notion of conventional cab signals out of your mind. That's not how this will work.

Yeah, conventional CSS is reliable, this new stuff won't be. Every time this sort of thing has gone in on transit it's ruined the system involved. The tradeoff is between a few accidents and getting where you want to go on time. Anyone who hates railroads is going to love a system that saddles them with an expensive burden and ties the system in knots while Joe the Trucker and the Bolt Bus fly down interstate 80 at 80mph door to door.

[Nasadowsk]

Railroads need to go with something reliable like the French TMV or KVB that either uses coded track circuits or track mounted data beacons (oh that's what ACSES is).

Bah, LZB's a better system.

Every time this sort of thing has gone in on transit it's ruined the system involved.

Works on the Canarse line, and it'd work a lot better if the NYCTA would get everything equipped on the line.

In any case, the 10 or so trains a day the average main line in the US sees, a cab signal system isn't going to even have any effect on running times. There's CSS equipped lines with 10 times the traffic out there in the world that work just fine.

[JimBoylan]

the problem will be that a computer model will not run a train as fast as an engineers judgment, either because of conservative margins for error or intermittent updates.

Port Authority Transit Corp.'s thoughts on the Lindenwold, N.J. High Speed Line were that a human operator is often more conservative than the Automatic Train Operation because the human might practice "defensive driving" to avoid being fired if the train doesn't stop in as short a distance as the computer thinks it should. Going down hill, a Lindenwold train coasts and increses its speed until it exceeds the signal system's speed limit, when the brakes automatically slow it down and then release. A human must run slightly slower to keep the train under the speed limit at all times, or possibly have to explain why not, at the risk of his job.
Washington Metro is another example. It's Automatic Train Operation just applies the brakes at a predetermined point (based on speed, location, and other factors) without worrying about overshooting the stopping point and hitting another train.
Are you suggesting that Postive Train Control computers will be programmed to run trains slower than past practice?