How "green" are trains, really? ... with DATA!!

Amtrak Unlimited Discussion Forum

Help Support Amtrak Unlimited Discussion Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Joined
Apr 5, 2011
Messages
6,090
Location
Baltimore. MD
It's sort of an accepted in truth in the environmental community that the environmental performance of rail is superior to that of other transportation modes. However, there's not a lot a real data on this, and it's usually couched in terms of "energy efficiency" which is a little bit different than actual emission levels. (There's also the issue of the environmental footprint of rail infrastructure as opposed to roads, airports, canals, etc, and the claim that rail might be better at enabling dense walkable urban centers that would be a viable alternative to suburban sprawl with its extremely negative ecological footprint.)

Last January at the annual meeting of the Transportation Research board in Washington, I saw a poster which actually provided some real-world fuel consumption data for an Amtrak train, something I've never seen before. This was done by a team from the Civil Engineering Department at North Carolina State University, and their main goal was to see if they could model energy consumption of various run segments and perhaps be able to demonstrate that "eco-driving" techniques could result in significant reductions in fuel consumption. As fuel consumption is very strongly correlated to CO2 emissions and reasonably strongly correlated to NOx and PM emissions, their approach might be useful in evaluating the claim that trains are environmentally superior, at least from the standpoint of emissions.

Here's the poster. You should be able to zoom it out to read the whole thing:

20190115_095156.jpg

The test train was the Piedmont service between Raleigh and Charlotte, with a particular emphasis on the Raleigh-Cary segment. The first thing that jumped out at me was a comparison of the fuel consumption of the different segments:

20190115_095221.jpg

The fuel consumption of the Raleigh-Cary segment averaged about 5 kg/diesel file per mile (+/- 1 kg/mi), whereas the other segments between Cary and Charlotte averaged about 3 kg/mile (+/- 1 kg/mi) The Raleigh-Cary segment ran at 0.6 to 1 mile per gallon (mpg), and the Cary to Charlotte segments ran at 1 to 1.9 mpg. The Raleigh-Cary segment is short relative to the segments between Cary and Charlotte, and a larger percentage of the run is involved in acceleration. There are also a couple of curves between Raleigh and Charlotte that require the train to slow down and then speed up, which means more acceleration. As anyone who tries to "eco drive" their car or truck knows, the more times you accelerate and the harder you do it, the worse the fuel consumption.

To compare this to other modes, you have to know how may passengers the train carries. From somewhere on trainweb, I found that the consist of the Piedmonts are usually 2 - 4 coaches and a vending/baggage car. They have two types of coaches, one has 56 seats, the other has 66 seats. By the way, if I ride the Piedmont, how can I tell which coaches are the 56 seat coaches (which I think would have more legroom)? This means a maximum capacity of the train of 224 - 234 passengers. That means that on the Raleigh-Cary segment, the maximum possible fuel consumption is 134 - 264 passenger miles per gallon. On the Cary-Charlotte segments it's 224 - 500 passenger miles per gallon. The assumption is that the train is fully sold out, and, of course, the fuel consumption does not include was is needed for turning the trains, etc., as yard switchers don't run on fairy dust. In comparison with other transport modes, a single occupancy car with a fuel economy of 25 mpg, will get 25 passenger miles per gallon, and if there are 4 people riding in the car, it will get 100 passenger miles per gallon. I'm not sure what kind of fuel economy intercity motor coaches have now, but based on my work with 18 wheeler tractor trailers, I would expect they should be able to do 6 mpg. If a 50 seat bus is filled up, that would give them about 300 passenger miles per gallon.

There are implications about this that I will post later, as I I need to go right now.​
 
Thank you for the analysis. I wonder how the LD trains in the West would compare, with the Superliners' higher profile decreasing efficiency, and long distances between station stops increasing it. Given the choice between the relative comfort of a motor coach and a train, being as there doesn't seem to be a significant difference in efficiency, I'd take the train every time...

Of course, any type of transportation is more efficient per passenger mile when the full capacity is utilized.
 
Trains with Amtrak LD legroom weigh a lot more per passenger than other means of transport. For example, an empty A321neo weighs 110,500 lb and carries up to 240 passengers. An empty J4500 weighs 36,300 lb and carries up to 60 passengers. Meanwhile, an empty Superliner I weighs 157,000 lb and carries only 75 passengers. That's excluding the locomotive and non-revenue service cars, whereas the A321neo and J4500 both include power.

At such great weight, physics make it impossible for a Superliner I to be more efficient than a J4500. At best it may equal an A321neo, and even that is dubious because the Superliner carries far less passengers than an A321neo while weighing far more even without locomotives.

Of course, in China a double-decker coach the size and weight of a Superliner would likely carry 180 to 200 passengers, defeating the efficiency of motorcoaches and airplanes. Then again, Chinese coaches feature 5-abreast non-reclining seats.

Speaking of the J4500, it gets about 6.6 MPG highway. My older, smaller 102D3 gets about 8 MPG highway and carries up to 47 passengers, though it's only equipped for 38 right now. However, my 102D3 gets less than 3 MPG city and I'd imagine a similar drop for other buses.

Keep in mind Amtrak's average load factor is 50%. The average Amtrak train is likely comparable in efficiency to the average Prius.
 
Last edited:
There are some limitations on the universal application of the analysis done by in to poster I cited. The most important is that the Piedmont is not 100% typical of all Amtrak trains (if anything is typical). First, Piedmont trains are pulled by FP59PH and FP59PHI locomotives, whereas most of the other trains in the Amtrak diesel fleet use Genesis or Chargers. It's not clear how these different locomotives compare in terms of thermodynamic efficiencies of the engines and emission performances, except that I would expect that the Chargers, being a more recent design, are superior in both respects. At an unscheduled "fresh air" stop (one of the engines broke down and we had to wait until some NS folks could come a throw a switch to a siding where they parked the broken engine) once on the Capitol, I perused the builders plate on the Genesis locomotive and found that in complied with EPA Tier 0. I don't think there was any EPA emission regulation of locos when the FP59PH's were built, and the Chargers comply with EPA Tier 4, which means that they're pretty clean.

The next is that the Piedmont has a pretty short consist for an Amtrak train. The poster said the trains they tested had only 3 cars. The Northeast Regional trains have 6 Amfleet 1 coaches at 72 seats each and a business class of 66 seats for a maximum total capacity per train of 498. During some busy times the Northeast Regionals might even have 7 or 8 coaches. Superliner coaches hold 75 passengers, Amfleet 2 coaches hold 60 passengers. Superliner sleepers can hold a maximum of 44 passengers, Viewliner sleepers can hold a maximum of 32 passengers. It's not clear what the effect of additional cars has on total train mpg. The additional weight should increase fuel consumption, but the additional passenger capacity should offset that when calculating fuel economy in passenger miles per gallon.

Of course, none of this analysis has anything to say about the fuel efficiency/emissions footprint of the electrically powered fleet. That would require separate testing and consideration of the mix of power generation sources for the juice used to power the NEC.

Probably the best way to measure the fuel efficiency/emissions footprint of something like Amtrak (or a bus line or an airline) is simply to take the total passenger miles and divide it by the total gallons of diesel fuel used. This would thus include the fuel used by the hard switchers and deadhead/non revenue moves. Fuel consumption should give a pretty consistent estimate of CO2 emissions, but the relation between NOx and PM emissions and fuel consumption will be different for the different emission tier locomotives, so evaluating emissions footprint from fuel consumption will be a bit more complicated.

Of course in Amtrak's case, this is further complicated by the fact that a good number of trains on the east coast run under both diesel and electric power (Silver Service, Carolinian, Crescent, Palmetto, Vermonter and all the Virginia trains, as well as the Empire service trains and Adirondack, Ethan Allen, and Lake Shore Limited.) In this case, not only do the passenger miles on those trains need to be allocated to diesel and electric portions, but the diesel fuel used by the yard switchers in Boston, New York, Philly and Washington will need to be allocated between electric and diesel operations. Thus, even the Boston-Washington Northeast Regionals have some diesel fuel emissions footprint, though probably very small. Oh, and if you want to be really picky, you should include the fuel used to ferry crew to relief points (especially if delays, etc. mean that crew changes are needed in unusual locations) and even the fuel used by the tanker trucks to travel to the refueling points. I suspect that such stuff represents a very small fraction of the total gallons per passenger mile, but I don't really know, so it would probably be a good idea to look at it more closely.
 
Trains with Amtrak LD legroom weigh a lot more per passenger than other means of transport. For example, an empty A321neo weighs 110,500 lb and carries up to 240 passengers. An empty J4500 weighs 36,300 lb and carries up to 60 passengers. Meanwhile, an empty Superliner I weighs 157,000 lb and carries only 75 passengers. That's excluding the locomotive and non-revenue service cars, whereas the A321neo and J4500 both include power.

At such great weight, physics make it impossible for a Superliner I to be more efficient than a J4500. At best it may equal an A321neo, and even that is dubious because the Superliner carries far less passengers than an A321neo while weighing far more even without locomotives.

Of course, in China a double-decker coach the size and weight of a Superliner would likely carry 180 to 200 passengers, defeating the efficiency of motorcoaches and airplanes. Then again, Chinese coaches feature 5-abreast non-reclining seats.

Speaking of the J4500, it gets about 6.6 MPG highway. My older, smaller 102D3 gets about 8 MPG highway and carries up to 47 passengers, though it's only equipped for 38 right now. However, my 102D3 gets less than 3 MPG city and I'd imagine a similar drop for other buses.

Keep in mind Amtrak's average load factor is 50%. The average Amtrak train is likely comparable in efficiency to the average Prius.

Vehicle mass is not the only factor determining power requirements and thus fuel consumption. An airplane not only needs power to move the vehicle, it also needs power to generate lift to keep the thing in the air. Power requirements for land vehicles like buses and trains are determined not only by the vehicle weight, but also by rolling resistance, aerodynamic drag, thermodynamic efficiency of the engine, grade, curves, acceleration, etc. Trains, even overweight American trains, have some advantages over highway vehicles with lower of coefficient rolling resistance (steel wheel on rail has coefficient of rolling resistance about a tenth of that of the typical truck/bus tire) and probably better aerodynamic performance, as a string of multiple rail cars doesn't generate the same total drag as each railcar punching through the air as a separate vehicles. Again, I'm not sure how much adding an additional railcar adds to the total fuel consumption of the entire train, but for sure, if you are running one bus and you need to add a second, you are doubling fuel consumption.That's why the truckers like to run double and triple trailers wherever they're legal.
 
Interesting. I went through this in the 1973-74 Energy Crisis for Oregon DOT. The Oakridge Laboratory produced a set of energy use comparisons that was widely-quoted by highway fans, using an EMD E-unit for the rail power. Aside from their assumption of 1950's technology, I found that the biggest differences were caused by assumptions about load factors, seating capacity, and stops. From recollection, a train with bus or air-density seating (3rd Class?) was most efficient, but even with their assumption that trains have lower density seating all of the public transport modes were more efficient than 1973 Detroit iron.

Some commenters here have alluded to the thing that we found to be most important: people who mainly rely on public transport by any mode tend to locate their homes and activities more efficiently than people who rely mainly on autos or the fly/drive hegemony. So we concentrated more on land use than on the slight differences between bus, rail and -- where appropriate -- air.
 
The fuel consumption of the Raleigh-Cary segment averaged about 5 kg/diesel file per mile (+/- 1 kg/mi), whereas the other segments between Cary and Charlotte averaged about 3 kg/mile (+/- 1 kg/mi) The Raleigh-Cary segment ran at 0.6 to 1 mile per gallon (mpg), and the Cary to Charlotte segments ran at 1 to 1.9 mpg. The Raleigh-Cary segment is short relative to the segments between Cary and Charlotte, and a larger percentage of the run is involved in acceleration. There are also a couple of curves between Raleigh and Charlotte that require the train to slow down and then speed up, which means more acceleration. As anyone who tries to "eco drive" their car or truck knows, the more times you accelerate and the harder you do it, the worse the fuel consumption.

We are occasionally reminded of the old adage, that if you want to go fast, don't go slow. In other words, working on the elimination of pinch points and slow sections can often return greater benefits to overall travel times than pushing for higher speeds on the sections where youre already going fast.

The above data shows that besides the travel times, elimiating mid-journey slow sections also has substantial eco benefits.
 

Amtrak's own Annual Report: https://www.amtrak.com/content/dam/...y-Performance-Report-September-2018-Final.pdf.

Average Load Factor at the bottom is 51%.

Vehicle mass is not the only factor determining power requirements and thus fuel consumption. An airplane not only needs power to move the vehicle, it also needs power to generate lift to keep the thing in the air. Power requirements for land vehicles like buses and trains are determined not only by the vehicle weight, but also by rolling resistance, aerodynamic drag, thermodynamic efficiency of the engine, grade, curves, acceleration, etc. Trains, even overweight American trains, have some advantages over highway vehicles with lower of coefficient rolling resistance (steel wheel on rail has coefficient of rolling resistance about a tenth of that of the typical truck/bus tire) and probably better aerodynamic performance, as a string of multiple rail cars doesn't generate the same total drag as each railcar punching through the air as a separate vehicles. Again, I'm not sure how much adding an additional railcar adds to the total fuel consumption of the entire train, but for sure, if you are running one bus and you need to add a second, you are doubling fuel consumption.That's why the truckers like to run double and triple trailers wherever they're legal.

Of course there's many other variables to consider with fuel efficiency. However, Amtrak trains' great weight is significant. The average car, bus, or airplane weighs about 500 lb per passenger. An Amtrak Western LD, including locomotives, Sightseer Lounge, baggage car, and Diner, weighs about 3000 lb per passenger. In contrast, the average train worldwide is likely about 1000 lb per passenger (at least the Chinese ones are, and since China has 20% the world's population and the world's largest rail network by passenger ASKs, that's a good place to start). No one can disregard this huge disparity in weight. A Western LD would need to be 6 times more efficient at moving a given unit of weight to match the efficiency of cars, buses, and airplanes. Such advantage is dubious at best. Just because train wheels have 90% less rolling resistance than Michelin XZA2 Energy doesn't mean it's 6 times more efficient at moving a given unit of weight.

As noted by Amtrak's own Annual Report above, Average Load Factor is 51%. In contrast, average airline load factor is above 80%. Unfortunately, your own MPG figures have such wide ranges that's it's hard to determine actual fuel burned per passenger-mile. Taking your figures and assuming average MPG is 1, average maximum capacity is 234, and Average Load Factor is 44% derived from Amtrak's Annual Report, the actual average PMPG is 103. That's bad, but consistent with the Piedmont's consist weight per passenger, adjusted for rolling resistance and drag.

According to actual operating figures, a 240-seat A321neo gets about 120 PMPG, depending on sector length. At 80% load factor, that'd be 96 actual PMPG. In other words, based on real-world data, the Piedmont is comparable in efficiency to the latest in airliner efficiency. That shouldn't come as a surprise since the Piedmont isn't the latest in rail efficiency, is overweight for its capacity, and has such poor Average Load Factor.

The latest J4500 gets 6.6 MPG highway with 60 seats for 396 PMPG. Even at Amtrak's average 51% load factor, it stills gets about 200 actual PMPG, and I'll bet the average J4500 has a lot more than 51% load factor. However, its fuel efficiency plummets below 3 MPG when operated in urban areas.

Please note, in China to achieve 234-seat capacity like the Piedmont, they'd use a 2-car DMU with 118 seats each. That DMU would weigh a hell of a lot less than the Piedmont, though with 5-abreast seating at 31" pitch (or with booth seating).

The most reasonable conclusion is that no form of transport is significantly "greener" than others. Any increase in maximum PMPG is often offset by poor load factors. Any increase in comfort is often offset by increased weight, drag, and maintenance cost (a 2-car DMU costs less to maintain than a Piedmont consist). Any reduction in rolling resistance, lift, or drag is often offset by speed or flexibility (planes fly faster than trains and buses go to more places more often on more direct routes). If I operated a J4500, I'll wager I'd have trouble filling all 60 seats all the time and would rather reduce seating to increase legroom. This isn't anything biased against trains or in favor of buses, but rather pointing out the facts based on real-world data. In fact, since buses are so inefficient in urban areas, the average transit bus is less efficient than a SUV and far less efficient than a Prius. Perhaps the real conclusion is that anything operated by the private sector (planes, intercity buses, and cars) are more efficient than their counterparts in the public sector.
 
Last edited:
Back
Top