When Nationwide Electrification is Called For
Small, dense developed countries should electrify their entire national rail networks. Usually, railroads think in terms of electrifying lines, but this hides the systemwide benefits of transitioning the entire network to run under electricity. I have previously written about this in the context of regionally funded commuter rail systems, as have Paul Druce and Clem Tillier. But some countries are so small and dense that the analysis for a single large metro area holds nationwide as well.
In this post I am going to focus on Israel, which is completely unelectrified, but also foray into mostly-electrified Belgium and the Netherlands, and currently-electrifying Denmark. Switzerland has already completed electrification; it is less dense than all of those countries except Denmark, but has cheap hydro power, which makes it cheaper to run trains under electricity, and key mainlines through mountainous terrain, where electrification is a major performance booster.
First, let us recall the performance benefits of electrification in flat terrain. The major rolling stock manufacturers sell DMUs with top speeds of 120-140 km/h, and EMUs with top speeds of 140-200 km/h; faster trains are generally more expensive, and with a few exceptions not of much use outside dedicated high-speed rail lines. The difference in acceleration performance is large: when the top speed is 100 km/h, an EMU such as the FLIRT takes less than 30 seconds to accelerate from standstill to top speed, corresponding to an acceleration time penalty of about 14 seconds, whereas the Stadler GTW DMU has a penalty of about 28 seconds (see data on PDF-p. 43); the GTW EMU version, a less powerful train than the FLIRT, loses 19 seconds. DMUs are also less comfortable than EMUs, because the diesel engines are right under passengers’ feet; longer-distance lines almost never use them, and instead use diesel locomotives, which accelerate even more slowly.
Because of this large difference in acceleration performance, electrification delivers the greatest performance benefits on lines with closely-spaced stops and high traffic. These are usually commuter rail lines rather than intercity lines. For example, suppose the top speed is 130 km/h, the stop spacing is 3 km, station dwell times are 30 seconds, and schedules are padded 7%. The FLIRT’s acceleration penalty is about 19 seconds, that of the diesel GTW (to 125 km/h) is 43 seconds; the deceleration penalties are both a bit lower than the acceleration penalties, but not too much lower, to avoid overheating. An EMU will average 68 km/h, a DMU 52 km/h. Independently of comparative energy and maintenance costs, this represents a 23% cut in the rolling stock requirement and in the on-board labor cost, and a larger cut in the required subsidy thanks to higher ridership. In contrast, if the stop spacing is 50 km, the difference in speed shrinks to 116 km/h vs. 113 km/h. Even if the EMU can do 160 km/h, its average speed is 140 km/h, still a smaller percentage difference than in the case of commuter rail, while the cost of providing this higher average speed is larger because tracks need to be upgraded to a higher top speed.
In small countries, short stop spacing is the normal state of affairs. In Israel, few segments of track have stops spaced more than 10 km apart, and those are mostly on the under-construction high-speed line from Tel Aviv to Jerusalem, which is planned to host 200 km/h electric trains. In the Tel Aviv and Haifa metro areas, stop spacing in the 3-4 km range is normal. Even intercity trains make all stops within Tel Aviv and Haifa proper, skipping the stations between those two cities. There are no major cities north of Haifa, only suburbs and small cities, and thus making many stops in and north of Haifa is justified for intercity trains – there aren’t many through-passengers who are being inconvenienced. South of Tel Aviv there are some moderate-size cities (as well as Jerusalem, but the legacy rail line to it is so curvy that the train from Tel Aviv takes twice as long as the bus), but because of high traffic, all trains make all four Tel Aviv stops.
With the exception of Belgium, all four countries under discussion also have dominant primate city regions, with about 40% of their respective national population; those city regions have dense rail networks, which are electrified in all countries except Israel. Denmark runs the Copenhagen commuter lines as a separate S-tog from the rest of the network, but in the Netherlands, Israel, and Belgium, there is no sharp difference. The result is that a large fraction of the overall rail network is urban commuter rail, which should be electrified, while additional chunks are regional rail with enough frequency to justify electrification even without a large city in the center.
Moreover, the service pattern makes it hard to electrify just a few lines in isolation, even if they’re the busiest. Regional rail networks frequently employ through-running. In small countries, this is common for the entire rail network, for different reasons: in Israel, the route through Tel Aviv is a new line from 20 years ago, without many platform tracks for terminating trains, whereas in the Netherlands and Belgium it’s the result of a highly nonlinear population distribution, which favors a mesh of lines, such that busy routes share tracks extensively with less busy ones. Compare these population distributions with that of the Northeastern US, where there is clear division into a trunk from Washington to Boston and branches heading inland.
Finally, these are all small countries. This is why I am not including South Korea in this proposal, even though it is denser, more mountainous, and more primate city-centric than all countries under discussion: South Korea is large enough that it’s plausible to run the Seoul-area commuter rail as an isolated electrified system, keeping the remainder of the legacy network unelectrified, with several maintenance shops for diesel trains around the country. In contrast, the unelectrified portion of the Dutch rail network consists of isolated branch lines, making it less economic to keep operating diesel trains. Israel has no electrification at all, but if it electrifies the Tel Aviv and Haifa commuter trains, the remainder of the network will be disjointed, requiring inefficient solutions such as considerable deadheading, or regular runs of diesel trains under long stretches of catenary.
One example I keep harping on, which I got from The LIRR Today before its blackout, is the LIRR’s diesel runs. The LIRR is almost completely electrified, and its diesel branches see little service, especially at the easternmost end of Long Island. Between this and work rules that separate diesel and electric train crew, the crew on one of the diesel trains work 2.5 hours per workday, running a train once in one direction and deadheading the way back; this and the bespoke nature of diesel trains on the LIRR lead to high operating costs.
The situations in the countries in question are not as comical as on the LIRR, but there are bound to be inefficiencies in Belgium and the Netherlands, and soon to be Denmark, which is electrifying its main lines, which together with the S-tog are a majority of its network. In Israel, the situation is the worst, since its rail network is even smaller: 1,100 km, compared with 2,600 km in Denmark, 3,600 in Belgium, and 2,900 in the Netherlands; this means that a partially electrified situation involves even smaller train orders and higher operating costs, while an entirely unelectrified network involves poor service in the urban areas.
Israel also has no rail links with any of its neighbors, nor any plans to construct any. This means that its branch lines are truly isolated, unlike those of the Netherlands, Belgium, and Denmark, which sometimes connect to other unelectrified lines in neighboring countries.
The way out of high diesel operating costs is to spend the money on completing electrification. As the example of Denmark shows, the costs are not outrageous: about $1.1 million per kilometer (I do not know whether track- or route-km, but I believe this is track-km). In the case of Israel, whose rail network is almost entirely single-track, this is not much more than $1 billion either way; to put things in perspective, the projected cost of the first Tel Aviv subway line is now up to $4.2 billion, while the Ministry of Transportation’s overall budget is $3 billion per year (PDF-p. 10), mostly spent on roads, in a country with only 300 cars per 1,000 people.
All-diesel railroads resist electrifying their busiest lines because they prefer to be able to let every train substitute for any train, and, for smaller operations, maintain all trains in one yard. For the same reason, small railroads with high traffic, such as the national railroads of dense countries, should instead go all-electric, in order to retain the benefits of interchangeable trains and maintenance facilities while also capturing the benefits of electrification. It’s not terribly relevant to the countries I’ve recently lived in, but for the same reason Switzerland fully electrified, similar small, dense countries should do the same.
Pedestrian Observations 
http://www.bane.dk/visArtikel.asp?artikelID=19670
Electrification contract is for 1330 track kilometers.
Can a case be made for fully-electric locomotives pulling DMU trains in inner, electrified areas? Would the increased acceleration on these segments make it worthwhile?
It wouldn’t increase acceleration, because loco-hauled trains have low initial acceleration. The initial acceleration is limited by adhesion, which is limited by the proportion of the train’s mass that rests on driven axles. This is how EMUs can have initial acceleration rates of 1 m/s^2, even low-powered subway trains, whereas loco-hauled trains can’t, even very high-powered ones like the TGVs.
For a hybrid solution, dual-mode MUs are much more suitable. The problem is that being able to run under both diesel and catenary raises procurement and maintenance costs, so they’re uncommon; countries and regions instead make passengers connect from a DMU branch line to an EMU main line.
What surprises me is why there aren’t any EMU trains that operate like railroad slugs, being able to take power off a coupled locomotive for their own traction motors in non-electrified territory. You’d have all the benefits of EMU on electric territory, and many of the benefits of DMU, without the unicorn dual mode functionality away from the wires.
There’s actually a reason for that: it means you need to somehow supply the power from the locomotive to the cars. And since the amount of power involved is quite large, you’re going to need either high voltage or high current, both of which have problems. Since you’re probably not going to put in a huge AC transformer on the diesel locomotive, and high voltage couplings are tricky to connect and disconnect, that’s not a great option, unless you have semi-permanently coupled sets with a dual-mode locomotive that already has a transformer. Lower voltage DC already exists on any diesel-electric locomotive, so that is a better option, but the current is potentially quite high, so you need big fat wires: a 750 kW EMU takes 1000 amps at 750VDC, which means you need a very beefy cable to supply a whole train: again, more doable with a semi-permanently coupled set.
The other way round may be more of an option. In fact, the SNCF had some TGVs pulled productively by diesel engines (connection to Caën until the line got electrified).
And Les Sables d’Olonne too (unless you’re confusing something).
Tangential observation:
During the WWII some Swiss steam-powered shunters were converted to electric!
http://www.aqpl43.dsl.pipex.com/MUSEUM/LOCOLOCO/swisselec/swisselc.htm
“Some” stands for “two” in this case. They were relaively successful, as far as I remember reading, but the conversion was apparently not worthy to be done on a bigger scale, and after the war, they were built back to standard. And if I remember reading correctly, one of them still exists on a museum line.
In the Netherlands there was recently a ‘business case’ study about electrification of the rest of the network. The official result is that LNG was a better option. Problem is that the study is textbook example of malice or stupidity. Par example the timetables were exactly the same for EMU & DMU’s, so no reduced material or labor. An increase of more travelers because of trip time reduction or quality was also discarded. And many more faults.
The link I gave to the GTW performance came from such a study. Annoyingly, the study compared the diesel GTW with the electric GTW and not with higher-performance EMUs; that’s why I bring up the FLIRT in this post.
FWIW, for practical reasons… A GTW could be electrified for a very low cost, because “only” the power cube had to be replaced, and the control software updated; If there were modifications in the cabs, they would be minor.
Of course, FLIRTs would be another league of train. Compared to a GTW 2/6, the capacity would be doubled or more. If the more powerful trains could generate sufficient additional traffic would have to be considered on a case by case basis.
The Milwaukee Road operated an electrified line over 600 miles from the Rockies to Seattle for much of the twentieth century. Operating costs were low, but the MBA accountants decided to switch to diesel in 1972, and soon after the Railroad went bankrupt. That part of the line is now hiking trails. Cost of fuel is now five times higher (adjusted for inflation), so converting most any line to electric seems like it would make economic sense.
The Milwaukee is a good example of how partially-electrified lines fall through the cracks, actually. It was electrified in two separate segments: over the Rockies, and over the Cascades. In the early 1970s, keeping the system as it was was not an option – the only options were to bridge the two electric districts, and to deelectrify. The railroad chose wrong.
It’s worth noting that corporate bond rates in the 1970’s were as much as 12 TIMES higher than they are today. Before the oil crisis, the operating improvements from electrification may not have ever been able to cover the debt service costs they created. Electrification would have been more resilient in the face of an increase in the cost of energy, which just about nobody reasonably had expected to occur.
The legends on railroad.net etc. are that management was cooking the books to make it appear that running electric trains was much more expensive than it was.
General Electric was willing to finance renovation of the system.
Milwaukee Road management had many problems. The ICC discovered (too late) that costs were being double booked. The railroad turned away $65 million worth of business for lack of car supply. The underlying problem seemed to be a tendency of management to think they had the answers, believing their own narrative instead of gathering objective information. Maybe in an age of spreadsheets and so on, creating the numbers would have been easier.
If someone is cooking the books your spreadsheet is going to have cooked numbers in it.
Spreadsheets didn’t spring full blown onto the market. People had been doing that sort of thing with ledgers and adding machines. They still sell the specially ruled paper so you can do it manually. The revolution with Visicalc and then Lotus 1-2-3 was that you didn’t have to cajole the high priest in the computer department to produce a report for you. Or break out some fresh columnar pads and the adding machine.
Nobody seems to know why the Milwaukee Road was double-booking costs on the “Pacific Extension” (for decades!) — it was only discovered after the bankruptcy, and after the Pacific Extension had been ripped up, during an audit.
The bad accounting led to bad decisions.
Wasn’t the Milwaukee’s transcontinental services the *only* part of the railroad that made any money? Their management is almost a textbook case of what not to do, because they let the very thing that brought them income—their permanent way and rollingstock—get run down and thus they were unable to provide a service, let alone break even or make a profit. Then there was the whole fiasco with selling off their cars and leasing them back at a higher rate. What Kool-aid were they drinking?
Not important to the argument but maybe interesting: If you subtract the uninhabitable area (The Alps) Switzerland is as densely populated as the Benelux.
Are there that many Swiss “key mainlines through mountainous terrain”? I can only think of the trans-alpine Gotthard and Lötschberg, both of which have been flattened by base tunnels. All others appear to stick to valleys and planes. What am I missing?
Switzerland fully electrified even before it built the base tunnels.
The Lötschberg route was electrified 15 kV/16.7 Hz from the beginning, 1913; the connecting Simplon route was first electrified in the Italian 3-phase system, and later converted to 15kV/16.7 Hz down to Domodossola.
The main reason for the electrification decision was the severe shortage of coal during World War I. The system decision was made in 1912, and after World War I, a serious round of electrification began; by 1928, the main lines were all electrified. Secondary lines were still steam operated, but World War II caused another serious shortage of coal, so that some lines got an emergency electrification, which covered secondary lines. In 1960, the last lines of the SBB network were electrified.
The other important alpine network, the Rhätische Bahn had their first electric line in 1913, and by 1922, their whole network was electrified.
Several other lines were built as what we now would call light rail / interurban streetcar standards, and electrified from the beginning (such as MOB, aka Golden Pass route, or the Berninabahn).
One thing to keep in mind about Swiss railroad lines… “flat” lines may have grades up to 2%. Of course, with electric traction, this is not such a big deal…
Point taken. Very interesting. Thanks, both of you.
A few thoughts on Denmark as I live here.
A lot of these transport decisions concern energy and balance of payments. Oil transformed the economy postwar and by the first oil shock 90% of energy consumption was oil (yes it was burned for electricity). This massive shock if anything put off electrification as the now very limited funds went into highways and bridges, with only electric suburban trains around Copenhagen receiving new investment. Partial electrification was attempted in the 1980s but was never finished as oil prices declined, with funding returning to road construction. By the late 1990s the state of rail was pitiful as it had not received adequate investments since the 1960s if not earlier. A decision was made not to electrify but instead buy a one-off diesel set from Ansaldobreda that could “offer the same benefits”. That has been such a monumental failure that it still is not in anything close to full service. Recognition of these issues politically have led to investment in electrification, new ERMTS signaling, a mix of upgraded rail and new 250 km/h double track, and switching to all electric trains on national lines. These began by mid-2000s and will last 25 years. Somewhat connected are investments in light rail in all regional cities / suburban Copenhagen (2010s) and the Copenhagen Metro (1990s-2020s) though they are more about urban development. To me at least it seems that 1960-2000 was about building out a highway network and the large bridges that connect the country to fit with an oil/suburban development pattern. As this is mostly complete investments are more easily directed to rail, especially as voters on the right will also benefit. As industry has left urban cores are attracting right leaning voters. In between cities suburban voters will benefit as rail lines are partly to move from their mid to late 1800s placements, to now follow the highways where regional development is concentrating anyhow.
Electrification in Denmark should be seen as a part of a much larger move in public investment from roads to rail, from suburban to urban, from a mostly complete infrastructure to an ancient one in desperate need of redevelopment.