US rail transport structure questions

[ac6000cw]

What may happen is that a railroad with long flat stretches and a few hills may accept slow running on the hilly sections as more cost-effective than using more locomotives over the entire journey or attaching and detaching helpers. I've been on Amtrak trains that did this too. That was probably one of the major impetuses behind adopting creep control and AC traction motors to increase tractive effort - the extra adhesion may mean less power is needed on the hills (accepting some slow running), so it may be possible to reduce the number/power of locomotives to just what is needed to keep a good speed on the flat. Also of course hilly sections are often tightly curved and unsuitable for running fast anyway

Yes, exactly, and it's why Burlington Northern kick-started the North American AC traction drive revolution in the the early 90's. Having invested an almost 'bet the company' amount of money in building a new line into the Powder River Basin coalfield in Wyoming, BN was later forced into sharing it with C&NW (effectively UP, although they weren't merged at the time). With a serious competitor on hand, lowering the operating costs for the (pretty long haul) coal traffic was essential - using multiple SD40s and C-30s with manned helpers to get them out of the coalfield worked but it was expensive. EMD with Siemens as electrical partner had been experimenting with AC-drives for a few years, so BN and EMD got together, EMD built four SD60MAC demonstrators and they strutted their stuff on BN. So impressed were BN that they ordered hundreds of the production SD70MAC version straight off the drawing board - the rest is history...(mostly GE's history, but that's a different story).

Three/four/five SD70MAC, SD70ACe, AC4400CW, ES44AC etc. have enough power to keep a heavy-haul train rolling at a reasonable speed on the flat, combined with prodigious continuous tractive effort all the way down to walking pace and below. Provided climbing the worst hills at 10mph or less doesn't eat up too much route capacity, they've become the North American heavy-haul loco of choice. Canadian Pacific's route west of Calgary basically doesn't use anything else for mainline freights.

Anyway, for a bit of weekend light relief, I came across some video from 2014 I recorded at Cassandra, PA on the Norfolk Southern (ex-PRR, ex-Conrail) line between Altoona and Johnstown. This is a roughly 1% eastbound climb towards the summit at Gallitzin (on the other side of the mountains is the famed Horseshoe Curve). It is roughly seven minutes of continuous trains, one uphill and two down - I'll leave others to decide if the trains are long, slow and infrequent or not... (for avoidance of doubt and for the freight car counters, there is roughly 30s of time missing just before the second train appears, otherwise it's the whole sequence in real time).

You even get three pairs of SD40E's as rear (manned) helper sets!

 

[eldomtom2]

I'll leave others to decide if the trains are long, slow and infrequent or not...

Long? Definitely. Slow? Well I can't say judging speed from videos is a particular skill. Infrequent? Impossible to tell from a single video.

 

[DelW]

Long? Definitely. Slow? Well I can't say judging speed from videos is a particular skill. Infrequent? Impossible to tell from a single video.

If you want to make your own assessment, you could try the Virtual Railfan channel on YouTube, which has cameras across the US rail system offering constant live streaming.

Cajon, Flagstaff or La Plata all have fast freight passing reasonably frequently on bi-directional double tracks. Though as with all US freight trains, there are no published schedules and there can be periods without trains due to track maintenance.

Ashland has a mix of (relatively) frequent Amtrak services among frequent freights. The route is bi-di double track through the town, but triple outside it.

 

[ac6000cw]

Well I can't say judging speed from videos is a particular skill

Easiest way is to count the freight cars over time. The longest single cars are usually stuff like Autoracks at around 85 feet, intermodal well cars probably average 50-60 feet (the longest container length is a 53 foot 'domestic' version).

Trains climbing decent gradients e.g. Cajon Pass average around 15 mph (I've paced some by road at that speed) - bulk trains can be slower, hot-shot intermodals faster. Once intermodals get moving on the reasonably flat and straight it can be hard to impossible to keep pace with them with a 55mph single-carriageway highway speed limit...

 

[eldomtom2]

Cajon, Flagstaff or La Plata all have fast freight passing reasonably frequently on bi-directional double tracks.

Of course those are the very creme de la creme of frequent freight in the US, existing as they do on the mainest of main lines used by a large variety of services, and in Cajon's case being on a bottleneck on which most freight into and out of one of the largest container ports in the world travels. Even then we're talking at most about four trains an hour - hardly the WCML. And of course looking at the number of trains passing a single point hides the frequencies of actual services. From a business perspective you want to minimise how much infrastructure you have to maintain and thus want to route as many trains as possible onto the same line.

And for reference, my reference to "long, slow, and infrequent" also applied to the railroads' approach to passenger traffic..

 

[ac6000cw]

Even then we're talking at most about four trains an hour - hardly the WCML.

It would be if the containers were being carried on UK length trains - the maximum length of a UK train is 775m, so subtract the loco length and it's around 750m of load length. That's 0.47 miles, which if you multiply it by 3 is about 1.4 miles - probably around a typical length for a double-stack train on that route (some will be longer, some shorter).

So each double-stack is carrying around six times the containers of a max length UK container train, and if the WCML was carrying the Cajon Pass traffic it would mean 24 freight trains per hour, every hour (or 576 per day).

From a business perspective you want to minimise how much infrastructure you have to maintain and thus want to route as many trains as possible onto the same line.

Yes, of course you do (it's self evident, isn't it?). It's why US rail operational methods have evolved over many decades towards long freight trains at a frequency sufficient to handle the traffic. It minimises crew and infrastructure costs but increases average transit times, but that's not so important for some traffic, especially if rail is the cheapest mode.

 

[eldomtom2]

It's why US rail operational methods have evolved over many decades towards long freight trains at a frequency sufficient to handle the traffic. It minimises crew and infrastructure costs but increases average transit times, but that's not so important for some traffic, especially if rail is the cheapest mode.

But that's exactly what I'm arguing! Passenger trains don't fit into that model.

 

[edwin_m]

It would be if the containers were being carried on UK length trains - the maximum length of a UK train is 775m, so subtract the loco length and it's around 750m of load length. That's 0.47 miles, which if you multiply it by 3 is about 1.4 miles - probably around a typical length for a double-stack train on that route (some will be longer, some shorter).

So each double-stack is carrying around six times the containers of a max length UK container train, and if the WCML was carrying the Cajon Pass traffic it would mean 24 freight trains per hour, every hour (or 576 per day).


Yes, of course you do (it's self evident, isn't it?). It's why US rail operational methods have evolved over many decades towards long freight trains at a frequency sufficient to handle the traffic. It minimises crew and infrastructure costs but increases average transit times, but that's not so important for some traffic, especially if rail is the cheapest mode.

Double stack doesn't double the capacity for the same train length. They are carried in pocket wagons so the length of the bogies is "dead space", although they are articulated so that's not as much length as it would be with conventional wagons. Also there are some restrictions on loading, such as not trying to balance a 40ft container on top of a 20ft, which may reduce capacity further depending on the mix of container types on a particular service.

The economics of freight transit change according to journey length. A transcontinental journey is several days so it doesn't matter so much if the container has to wait longer for the next train than it would in the UK where a transit is measured in hours. Even if all the many reasons preventing operation of long and/or double-stack trains in the UK were suddenly overcome, there would be few if any trains actually able to take advantage.

 

[ac6000cw]

Passenger trains don't fit into that model.

Yes they can and do - if you carry on west on BNSF from the bottom of Cajon Pass (at San Bernardino) towards downtown LA via Riverside, there are Metrolink and Amtrak passenger trains sharing the route in increasing numbers as you get closer to LA, particularly west of Fullerton. Some of the freight is for terminals at Commerce, the rest peels off in the Redondo Junction area to head south to the ports in the Long Beach area. The passenger traffic continues on to LA Union Station.

The Amtrak regional services that run between Oakland/Sacramento and Bakersfield use the basically single-track BNSF route down the Central Valley, shared with the BNSF freights that travel over Tehachapi Pass.

Double stack doesn't double the capacity for the same train length. They are carried in pocket wagons so the length of the bogies is "dead space", although they are articulated so that's not as much length as it would be with conventional wagons.

Yes, I know and I just forgot to factor it in - apologies (They sometimes do things like putting 40' maritime or 45' domestic containers in the bottom of a 40'/45' well car with 53' domestic on top to make some use of the dead space over the bogies).

If you allow for the dead space (which we have some of too, more so with some low-height well wagons) I think the ratio would be between four and five times as many UK-length trains - 16 tph/384 tpd is still quite a lot.

Even if all the many reasons preventing operation of long and/or double-stack trains in the UK were suddenly overcome, there would be few if any trains actually able to take advantage.

I agree - based on casual observation at places like the Felixstowe branch and Ely, intermodal trains vary in length and/or are not fully loaded sometimes. Plus most of them seem run direct from origin to destination, without any intermediate 'block swapping' which can happen in the US on the long hauls.

 

[eldomtom2]

Yes they can and do - if you carry on west on BNSF from the bottom of Cajon Pass (at San Bernardino) towards downtown LA via Riverside, there are Metrolink and Amtrak passenger trains sharing the route in increasing numbers as you get closer to LA, particularly west of Fullerton. Some of the freight is for terminals at Commerce, the rest peels off in the Redondo Junction area to head south to the ports in the Long Beach area. The passenger traffic continues on to LA Union Station.

The Amtrak regional services that run between Oakland/Sacramento and Bakersfield use the basically single-track BNSF route down the Central Valley, shared with the BNSF freights that travel over Tehachapi Pass.

Well yes, you can combine them if you get taxpayers to pay for all the necessary capacity improvements and all the costs of operating them. And when those passenger trains still operate at one train per hour frequencies at best.

 

[ac6000cw]

Well yes, you can combine them if you get taxpayers to pay for all the necessary capacity improvements and all the costs of operating them.

Is there a problem with that?

And when those passenger trains still operate at one train per hour frequencies at best.

I agree that the San Joaquin trains are not especially frequent (6 trains per day each way currently), but it's a much busier route between Fullerton and downtown LA with two Metrolink routes and Amtrak Surfliners sharing the tracks with freight.

 

[eldomtom2]

Is there a problem with that?

No, but it illustrates my point.

but it's a much busier route between Fullerton and downtown LA with two Metrolink routes and Amtrak Surfliners sharing the tracks with freight.

Even then we're only talking about 27 trains a day on the busiest section, with minimal service out of LA in the mornings and into LA in the afternoons.

 

[ac6000cw]

No, but it illustrates my point.

Err, which point?

with minimal service out of LA in the mornings and into LA in the afternoons.

Yes, I know, but that's typical of many US commuter rail services anyway.

 

[eldomtom2]

Err, which point?

That passenger traffic didn't fit with the model US railroads adopted.

Yes, I know, but that's typical of many US commuter rail services anyway.

I don't think this aspect of US commuter rail is separate from them running on freight railroad trackage.

 

[Taunton]

And are there currently any privately operated passenger services in the US think Brightline is one is this the only privately operated passenger service is the US was just thinking this as I saw an interview with Paul Krugman who is a liberal economist was critical of rail privatisation in the UK and he said 'not even the US has privatised passenger rail' just asking is this true but you could say it is not true of the infrastructure outside the NE corridor which is owned by the freight companies.

It can be the same. The UK approach is essentially private companies doing the actual operation as contractors on a public system. There are a number of instances of exactly the same in the USA - Herzog for example as operating contractor to various public agencies across the country. Herzog also do infrastructure maintenance work as a contractor for those tracks which are owned by the agency. There are various combinations.

And when Amtrak first started, many of their services were contracted back to the onetime rail operator on the route to actually run it.

 

[edwin_m]

That passenger traffic didn't fit with the model US railroads adopted.

I don't think this aspect of US commuter rail is separate from them running on freight railroad trackage.

Does anyone know the track access charging structure the commuter railroads use? In the UK we have a fixed charge that reflects the cost of providing the infrastructure, and a relatively low variable charge reflecting the extra cost of running an extra train on it. This incentivises operators to run more trains, and the other costs of providing an extra off-peak service are quite low because they use of trains and often crews that are needed anyway to provide the peak service.

If the US operators have to pay a much higher fee per train then they are likely to run trains only at the busiest times, which might explain why many of them just run a few into the city in the morning and out again in the late afternoon, despite this being a very poor use of the assets. Alternatively it could be a cultural thing embodied in the very term "commuter rail" - the assumption is that everyone drives everywhere by default and the train is used only when the roads are too busy. Another factor is that the owning freight railroad may demand access for their own services at other times of day.

There are exceptions that provide a tolerable off-peak service, such as the Long Island and Metro North - but it may be pertinent that both of these own their track.

 

[Jan]

There are exceptions that provide a tolerable off-peak service, such as the Long Island and Metro North

Even for those the "tolerable" aspect is debatable. Looking at the suburban service, on Metro North only the New Haven Line is really half-hourly (but skips almost all of the intra-city stops in the Bronx), the Harlem Line already has a few hourly gaps and on the Hudson Line only some stations get a second train per hour off-peak, and even that with noticeable gaps around mid-day. The LIRR isn't really better, either, with quite a few of even the electrified inner suburban branches only running hourly off-peak.

 

[ac6000cw]

I suspect the 'mostly into the city in the morning, mostly out of it in the evening' nature of a lot of US 'commuter rail' services is due to a mixture of history and politics (i.e. the funding for it). There always seems to be a fair amount of opposition to subsidising public transport, so there's constant pressure to minimise the operating costs and to not be seen to be running 'empty' trains (as I suspect doing that provides easy political ammunition to the opposition). So any 'reverse commute' trains I suspect are mostly just cycling trainsets back for another 'forward commute' trip.

Also the 'peak' service hours can be quite long e.g. the LA Metrolink 'Antelope Valley' morning inbound service starts with the 03:41 from Lancaster and ends with the 09:11, then the afternoon/evening outbounds from LA start at 15:39 and finish at 21:39 (with three 'reverse commute' trains in both peaks and a few shorter workings). AFAIK the majority of the infrastructure Metrolink runs on is owned by local/state authorities and dispatched by Metrolink (with some mostly local freight running over it on trackage rights).