News   GLOBAL  |  Apr 02, 2020
 8.4K     0 
News   GLOBAL  |  Apr 01, 2020
 39K     0 
News   GLOBAL  |  Apr 01, 2020
 4.7K     0 

PTC and ATC are part of the spectrum from ATP to ATO, which describe different levels of automation.

ATP is Automatic Train Protection. This is just a traditional signal system which monitors the train and intervenes if it violates any signal aspects (e.g. overspeed, passing red signal). I'm not familiar with the American PTC standard, but I believe it is a form of ATP.
PTC is a sort of form of ATP, but installed as an overlay on whatever existing system is already there rather than integrated into the signal system. It doesn't meet the traditional definition of ATP because of that, but still performs largely the same functions.

It should be noted that neither PTC nor ATP are by themselves signal systems. PTC is an installation on an existing signal system, while ATP is a function that can be built into it.

It should also be noted that there does exist at least one signal system in the North American sphere of operation which meets a lot of what people envision as the requirements. Amtrak's ACSES isn't a moving block system, but allows for virtually all of the same functionality of an ETCS Level 2+ installation.

Dan
 
Wikipedia says "Originally, Caltrain employed Parsons Transportation to develop a custom PTC system, called CBOSS, for CalMod, but due to delays, Caltrain switched to Wabtec and their I-ETMS system."

Everything to do with California rail expansions seems to end in fiasco, largely due to repeated attempts to reinvent the wheel. From what little we've seen so far, it looks like DB and Alstom are playing it safe by proposing proven equipment and technologies, which should save us headaches in the future, as long as Transport Canada doesn't force them to adapt systems to meet the arbitrary differences between TC regulations and EU regulations.

My biggest question is whether Alstom will propose ETCS Level 2 or Level 3. Level 2 has been in service for decades on many lines across Europe, but it doesn't support moving blocks. Level 3 can theoretically provide up to 30 tph on a single track, but it's less proven - when I wrote that assignment there weren't any systems fully in operation yet. There are also challenges related to its purely communication-based nature, primarily related to other trains which need to share the tracks.

In its purest form, ETCS L3 doesn't require any wayside detection equipment (e.g. axle counters, track circuits), which in theory makes it cheaper to maintain than L2 or L1. But in order to achieve that, all trains on the line need to have train integrity monitors. That means that the train itself needs to confirm that the train is still in one piece, and a coupler didn't break, leaving a car sitting on the line behind. This is fairly straightforward for passenger trains, but would be a large ask for freight operators given that freight cars roam around the continent, and therefore virtually all cars on the continent would need to be retrofitted. Although CN no longer owns the core rail network in Toronto, they do still have trackage rights, and do therefore have some influence on the decisions related to signal systems.

To overcome these limitations, ProRail (Dutch railway operator) has proposed to use an ETCS Hybrid L3 system where instead of moving blocks, they would use micro blocks. Each physical block would be divided into many virtual blocks, which would be used by (passenger) trains equipped with integrity monitoring. The (freight) trains which lack monitoring would only have access to the larger physical blocks, using the wayside integrity monitors. Given that moving blocks would only get updated every 10 seconds or so anyway, this provides nearly the same capacity for passenger trains as a moving-block L3 system, while reducing equipment requirements for freight operators. The fact that the micro-blocks match up with the physical blocks makes it much easier to overlay the physical and virtual movement authorities than with a system which combines moving blocks for passenger trains with fixed blocks for freight trains.

To cut on maintenance costs, ProRail also suggests that the physical block sizes could be increased compared to today, since the freight trains can be scheduled during times when line capacity is not critical. Doubling the physical block size cuts the amount of detection equipment in half.
Pretty sure Alstom was involved too . . . PTC ain't great . . .

ETCS L3 actually is already implemented in everyone's favourite railway https://en.m.wikipedia.org/wiki/Wuppertal_Schwebebahn
PTC is a sort of form of ATP, but installed as an overlay on whatever existing system is already there rather than integrated into the signal system. It doesn't meet the traditional definition of ATP because of that, but still performs largely the same functions.

It should be noted that neither PTC nor ATP are by themselves signal systems. PTC is an installation on an existing signal system, while ATP is a function that can be built into it.

It should also be noted that there does exist at least one signal system in the North American sphere of operation which meets a lot of what people envision as the requirements. Amtrak's ACSES isn't a moving block system, but allows for virtually all of the same functionality of an ETCS Level 2+ installation.

Dan
Yes, but obviously ACSES has a ton of problems, especially not being as well developed as ETCS
 
Japan Rail, especially for the urban lines is the gold standard when it comes to this matter. How do they do it?
 
My biggest question is whether Alstom will propose ETCS Level 2 or Level 3. Level 2 has been in service for decades on many lines across Europe, but it doesn't support moving blocks. Level 3 can theoretically provide up to 30 tph on a single track, but it's less proven - when I wrote that assignment there weren't any systems fully in operation yet. There are also challenges related to its purely communication-based nature, primarily related to other trains which need to share the tracks.

ETCS Level 2 can comfortably accommodate 3 minute frequencies if the block size is small enough; and thankfully the congested Union corridor doesn't need to handle large trains (just the occassional Canadian, no freight). That'll probably good enough for 20 years.

Also, IIRC, Level 3 doesn't support track-side equipment; every train requires a position broadcast mechanism. That makes it hard to support regular VIA Rail and freight services as they'd all be ghost trains requiring manual clearance.
 
Last edited:
Yes, but obviously ACSES has a ton of problems, especially not being as well developed as ETCS
I'm not so sure it's that obvious.

It did go through a protracted development, sure. But it's been in regular use for several years, and issues with on-board equipment have been non-existent. And hell, it was able to be upgraded to fully meet the PTC regulations well before the deadline.

ETCS Level 2 can comfortably accommodate 3 minute frequencies if the block size is small enough; and thankfully the congested Union corridor doesn't need to handle large trains (just the occassional Canadian, no freight). That'll probably good enough for 20 years.
Sure, but that's a function of the block size rather than the signal system itself. The Weston Sub can also handle sub-3-minute frequencies in its current state.

Dan
 
Also, IIRC, Level 3 doesn't support track-side equipment; every train requires a position broadcast mechanism. That makes it hard to support regular VIA Rail and freight services as they'd all be ghost trains requiring manual clearance.
I wouldn't say it "doesn't support"; rather, it "doesn't need". A train can operate in L3 on a line that is equipped with wayside signals and conventional train detection, but neither will be relevant to the operation of the train. You're correct in saying that every train needs a train integrity mechanism, i.e. it needs to positively assert that the back of the train is still attached to the front. (Trains in L2 and L3 have a "position broadcast mechanism", this isn't unique to L3.)

The requirement for train integrity is what makes operating in what some might call a "pure L3" wayside (with no train detection at all) difficult, mostly for freight. In the EU, the Digital Automatic Coupling (DAC) program is aiming to solve this problem by creating an end-to-end data path along the train that can assert integrity (among other things).
 
How Advanced Signalling Technology……

Good to see ML has set a direction. It’s hard to argue against this technology - clearly, it is proven technology that works.

The only big risk is that other key operators won’t follow, and we have an interoperability issue in those places where GO uses freight railroads. That can be overcome, I’m sure.

In this case, debating further merely creates delay. I hope this proceeds promptly.

- Paul
 
How Advanced Signalling Technology……

Good to see ML has set a direction. It’s hard to argue against this technology - clearly, it is proven technology that works.

The only big risk is that other key operators won’t follow, and we have an interoperability issue in those places where GO uses freight railroads. That can be overcome, I’m sure.

In this case, debating further merely creates delay. I hope this proceeds promptly.

- Paul
well for areas like the kitchener corridor, the trackside signaling stays in place. unless they want to upgrade themselves, but considering the frequency its unlikely.

I could see MX paying cn/cp to upgrade the signaling though
 
well for areas like the kitchener corridor, the trackside signaling stays in place. unless they want to upgrade themselves, but considering the frequency its unlikely.

I could see MX paying cn/cp to upgrade the signaling though

The issue would be whether CP/CN would accept the installation if it were other than what they intend to use for their own trains.

But yeah - new signalling only makes sense in the core areas where we are expecting more intensive, close-headway operation. Even 15 minute headways to Kitchener or Hamilton don't require it, although eventually TC may insist on some form of more advanced signalling for all passenger trains (at least).

- Paul
 
One advantage of having such primitive signalling technology is that it may be easier to implement the handoff where trains pass from the central segments with ETCS to the outer segments with our current signal system. Does our current system even have Automatic Train Protection?

Tons of railway projects have been delayed due to challenges getting ETCS signalling to properly connect to adjacent signalling systems. As we speak, the Crossrail project in London UK is scambling to iron out the bugs when trains cross between the ETCS signalling on the Great Western mainline, the CBTC signalling in the tunnel, and the TPWS signalling on the Great Eastern mainline. That's why the line will initially operate as three disconnected services, each staying within a single signalling system.
Knipsel.JPG

Knipsel1.JPG
 
Last edited:
The only big risk is that other key operators won’t follow, and we have an interoperability issue in those places where GO uses freight railroads. That can be overcome, I’m sure.
It's a big risk for sure, but not in the way you might think. There's no reason why CN's operations on GO's network need to change at all from how they are today. Given that ETCS's singular underlying design goal was interoperability, the system (writ large) is flexible enough to accommodate a wide variety of scenarios.

Where I see the risk is from Transport Canada's proposed Enhanced Train Control (ETC) mandate. In fact, a few pages of my comments to TC were on the topic of freight interoperability on passenger networks. Here's an excerpt:

5.1-comments-TC-ETC.png


I'm sure that ONxpress, Metrolinx and CN will work together to reach a mutually beneficial outcome, but the big unknown right now is the specific requirements that may be imposed by TC.

Tons of railway projects have been delayed due to challenges getting ETCS signalling to properly connect to adjacent signalling systems. As we speak, the Crossrail project in London UK is scambling to iron out the bugs when trains cross between the ETCS signalling on the Great Western mainline, the CBTC signalling in the tunnel, and the TPWS signalling on the Great Eastern mainline. That's why the line will initially operate as three disconnected services, each staying within a single signalling system.
I get what you're trying to say, but I'm not sure I agree with the framing. ETCS certainly has its issues, but the Crossrail issues have very little to do with ETCS specifically. Transitioning between level NTC-TPWS and levels 1/2/3 is solved: hundreds of trains do it every day in Britain. While it's regrettable that ETCS was not chosen for the core section of Crossrail (though it was a justifiable decision at the time), the majority of the signalling issues on Crossrail stem from software onboard the trains and functionality of the bespoke/proprietary CBTC system in the core. I recall reading lists of hundreds of deficiencies, some of which were real head-scratchers. I'm pleased they continue to make progress but it's been nothing short of a systems integration disaster.

Remember that ETCS must interface with -- but is fundamentally separate from -- the railway interlockings. In many cases, the "connection to adjacent signalling systems" doesn't change when ETCS is overlaid (by design). What does this mean for GO? The interfaces between Metrolinx's signalling plant and CN's signalling plant don't need to change much (or at all). GO trains will transition from level 0 to level 2 (potentially via level 1) at the boundary and back again with no involvement from CN. Maybe there will even be a future level NTC-IETMS, but we'll have to wait and see what Transport Canada does.
 
So this is the first I'm hearing that GO is using ETCS instead of PTC on the network. It sounds better overall, but I am wondering one thing: if freight in Canada installs traditional PTC, is there any ability of interoperability on the networks? Like, for the section on the Kitchener Line between Bramalea and Georgetown that is CN owned, would a GO train with ETCS be able to use PTC in that moment? Or would the train have to revert back to fixed block.

Also, let's say that CN puts PTC on their tracks and VIA adopts using PTC on their trains for movement on the CN tracks, when a VIA train comes onto a GO train track section (as they do into Union) will they be able to use ETCS or will they be in the dark as well and rely on backup fixed block signaling?

The same for Amtrak trains, its my understanding that the US is going all-in with PTC.

Would love some clarification thanks!
 
The answer is a great big "it depends". Every scenario you posted is technically possible, it all depends on what ends up get spec'd and installed.
 

Back
Top